Although AI-enabled interactive decision aids (IDAs) have demonstrated to provide reliable advice, users are rather reluctant to follow this advice. One recently highly discussed reason for this reluctance is users’ perceived unclear accountability of the AI-service regarding the decisions of these AI-based IDAs. Drawing on accountability theory, we designed user-interface (UI) design artifacts for AI-enabled IDAs based on the dimensions identifiability, expectation of evaluation, awareness of monitoring, and social presence and tested them through a scenario-based factorial survey method (N = 629). We show that accountability-emphasizing UI design artifacts individually raise users’ accountability perceptions of the AI-enabled service, which in turn influence users’ compliance to follow the advice from the AI-enabled service. These findings have important theoretical and practical implications, particularly as they inform how to increase the transparency of accountability of AI-enabled services and thus user compliance.
Adam, Martin, "Accountability-Based User Interface Design Artifacts and Their Implications for User Acceptance of AI-Enabled Services" (2022). ECIS 2022 Research Papers. 5.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.